      
|
||||||||||||
|
Center Administration |
 |
CENTER ADMINISTRATIONAdministrative Core
INTERNAL ADVISORY BOARDThe Internal Advisory Board is chaired by Dr. Federoff and consists of Paul Coleman, Ph.D.; T. Franklin Williams, M.D.; Robert Griggs, M.D.; William Hall, M.D.; and Robert Frisina, Ph.D. This board assists in policy decisions. The Dean of the School of Medicine and Dentistry, Deborah Cory-Slechta, Ph.D., serves as an ex-officio member of the board. The Internal Advisory Board meets quarterly and is particularly important in providing input related to new directions for the Shock Center.
EXTERNAL ADVISORY BOARDOversight of the External Advisory Board is carried out by T. Franklin Williams, M.D. who coordinates the annual review of the Nathan Shock Center by the External Advisory Board. An annual meeting of the External Advisory Board is convened to review all aspects of the Shock Center and to provide recommendations about potential new directions.
ANNOUNCEMENTS AND PUBLIC INFORMATIONSelected Abstracts of theRochester Nathan Shock Center Bioinformatics Workshop October 20-21, 2001
"Computational Approaches to Gene Expression
Data Analysis" Microarray technologies including high-oligonucleotide and cDNA arrays enabled monitoring of the mRNA levels of thousands of genes in a single experiment. Methods for analyzing the torrents of high-dimensional data are in need. Machine learning, an area of computer science, studies computational methods for extracting information from data. Application of machine learning techniques is thus a reasonable approach to the problem of analyzing gene expression data. This talk presents recent developments in the use of machine learning techniques for clustering and classifying gene expression data. The techniques to be covered include Self-Organizing Map and Support Vector Machine. The underlying principles and limitations of the techniques will be also discussed.
"Design, Application and Analysis of Membrane
Based cDNA Arrays" Recently, cDNA arrays have allowed the simultaneous analysis of gene expression patterns of thousands of genes in parallel. However, the basic hybridization theory underlying current cDNA array approaches has been available for over three decades. Current approaches to array use and analysis sometimes employ technological approaches or costs that limit the universal application of this technology. In developing a large scale cDNA array effort, we have attempted to use everyday molecular tools, laboratory equipment, computer hardware and software resources available to most laboratories whenever possible. This has included the design and use of nylon membrane based cDNA arrays. With this approach, we have produced over 17,000 cDNA arrays and have applied them in the analysis of complex systems including asthma, schizophrenia, drug abuse, and the analysis of model systems.
"Designer Microarray Approach: From Soup to
Nuts" We shall present a methodology of do-it-yourself, next-generation biochips. Recent advances in microarray technology have made the DNA chip approach a very popular means of unravelling the complex genetic traits governing a biological process. At present, many commercial DNA chips or microarrays are available to researchers, to decipher the molecular status involving hundreds or thousands of changes in gene expression. In general, these DNA chips are touted as "gene-screening" tools based on a platform of thousands of genes; often the results are presented as identifying a few hundred genes, showing qualitative or quantitative changes in terms of a few fold of either down- or up-regulated differences. From the published records of these types of results, it becomes apparent that there is the need to develop the next generation of biochips, allowing investigators to take data from macro-DNA chips based on platforms of 20,000 to 30,000 genes to the next step gene-screening task, zeroing in on identifying those few genes that are the keys to a particular biological process. I shall present a soup-to-nuts approach enabling the generation of medium-density biochips, describing the technology from how to make the biochips with genes of interest, or genes identified by the macro DNA chips, to how to perform adequate bioinformatic analysis, including tasks involved in data mining, data organization, and data security and archiving. I will also present essential considerations of "controls" and "standardization" from the biological, engineering, and software aspects for "true" and "false" in the biochip design and data mining steps. The entire approach developed in my laboratory can be easily reproduced in any molecular biology laboratory with standard laboratory equipment, with minimal investment in additional equipment or personnel. In addition, our designer microarray approach is not only inexpensive with do-it-yourself features, but also provides the means to perform "gene-hunting" with a divide-and-conquer approach, designing a cassette of biochips based on a series of selected functional or regulatory modalities controlling, in this case, the aging process. For example, one can use our approach to design IGF-1 signalling pathway biochips to define the concordance and discordance in gene expression related to longevity in different species, from mice to yeast. Similarly, one can use such pathway-based designer microarrays to study the tissue-specific profiling of gene expressions in mice during the aging process. Therefore, our designer DNA chip approach provides the technological next step for biologists in their quest for the "master genes" that control complex processes, and will ultimately provide biosignatures for the processes of normal aging, and signature gene patterns for age-dependent diseases.
"Designing Microarray Experiments to Answer
Biological Questions" There are two types of variability or noise in microarray experiments: technical and biological. To obtain valid answers to biological questions requires acknowledging and estimating biological variability, not just technical variability. Therefore microarray experiments should be designed to include biological variability and to allow estimation of its magnitude. Design principles to achieve this will be discussed, including issues related to use of replication, number of replicates (sample size), and pooling of biological samples.
"Examining microarray data for biological response"
Analysis of microarray data and assessment of the significance of results requires statistical approaches. The approaches taken to determine which genes show reproducible changes, and which biological processes are involved are discussed in the context of analysis of aging in C. elegans. The expression patterns during the normal aging process in C. elegans were profiled. Spotted DNA microarrays consisting of 89% of the C. elegans genes amplified from genomic DNA were used. We isolated RNA from synchronized cultures of sterile animals at various ages from young adult to old age. We hybridized labeled cDNA made from this RNA to DNA microarrays, and used one way ANOVA to identify 212 (p<.001) genes with changes in expression with chronological age. These aging genes were compared with functional sets of genes to look for significant over-representation or exclusion. A complementary technique, examining functional groups for concerted changes during the time course was also used to find age-dependent biological processes. The aging profile of global clusters of genes identified previously was also examined. The global clusters were identified by finding the correlation between each pair of genes followed by force-directed placement to identify co-regulated genes in 553 worm experiments. Using these techniques, we found transcriptional changes and biological processes involved in aging.
"High Density Oligonucleotide Arrays" Bioinformatics has emerged as a new field of biomedical research with the rapid proliferation of large-scale public and private databases of genomic, proteomic, and pharmacogenomic information. Managing and mining these massive archives of information for knowledge that can be used to generate new hypotheses for biological mechanisms and treatments for disease is a formidable challenge for both computer scientists and biologists. A further challenge is presented by the large quantities of detailed quantitative information currently generated by large-scale gene expression measurements on DNA microarrays and by biomedical imaging technologies. The analysis and mathematical modeling of "all these numbers" requires additional skills from biologists with strong analytical and computational skills, and computational scientists (eg. physicists, mathematicians, and engineers) with a commitment to biological problems. Several case studies using Affymetrix Genechips will be presented to illustrate some of the challenges of "separating signal from noise" and then linking the measured gene expression profiles to the vast array of biological and genomic databases.
"Preprocessing of cDNA microarray data: Normalization
and other issues" The raw data from cDNA microarray experiments are two image files of the slides. Reduction of these images to spot to foreground and background values is an important part of the preprocessing of this data, as are the various normalization tasks such as intensity-based and spatial color balancing within a slide, and scale equalization across slides. In this talk I will outline the issues here and describe some approaches to addressing them.
Production and Use of cDNA Microarrays in Studies
of Aging" The Buck InstituteŐs Gene Expression Resource Center is focused on supporting all the research investigators needs for gene expression profiling. The Center has developed and implemented methods and procedures for the production and use of DNA microarrays. Spotted microarray technology combined with the CenterŐs ability to quickly produce restricted clone sets has enabled the Center to produce a wide range of microarrays that are custom tailored to meet the individual researchers needs. The Gene Expression Resource Center designs and produces many different high-density microarrays from PCR amplified cDNA clones using our collection of over 75,000 sequence verified Unigene clones from the murine, human, Drosophila and C. elegans genomes. Using investigator provided RNA samples the Center labels the RNA, hybridizes the RNA to the microarrays, scans the hybridized microarrays, quantitates the scanned images, annotates the datasets and assists the research investigators in the experimental design and interpretation of their gene expression studies. The Center also offers laser capture microdissection (LCM) and linear amplification of RNA technologies to enable research investigators to isolate and measure gene expression from the individual cell types using DNA microarrays. To validate gene expression changes identified by microarray analysis the Center also offers resources for Quantitative Reverse Transcription PCR (Q-RT-PCR). Data will be presented on methods development, testing, use and analysis of cDNA microarrays in gene expression studies of aging.
"The Bio in Bioinformatics" The increasing use of large-scale genomic approaches in the analysis of complex biological systems has required the development of computational strategies in the analysis of genomic datasets. cDNA array analysis in particular has necessitated complex approaches to unraveling biological complexity. This generally takes two forms: a) purely computational or algorithm based approach or, b) annotation based approachs. We show examples of annotation based approaches to microarray and biological analysis to view gene expression data in the context of biological pathways as well as linkage and association data of complex human disease.
"The Design and Analysis of Microarray Experiments" Spotted cDNA microarrays are a powerful and cost-effective tool for high-throughput analysis of gene expression. As the potential of this technology has become apparent, it has also become obvious there are many important issues in experimental design and data analysis. There are many different sources of variation in microarray studies and the data are inherently "noisy." Sound statistical techniques are essential to compute unbiased estimates of relative expression from microarray data and to judge whether there is significant evidence of differential expression. However, whether this will be possible - whether the data contain the necessary information - depends on the experimental design. Notions of statistical inference can also be extended to evaluate the results of higher-order analytical tools, such as clustering.
"Use Of cDNA microarrays to examine changes
in gene expression in response to oxidative stress" Most bird species have a much greater maximum lifespan potential than mammals of similar size. However, the molecular mechanisms that birds use to achieve their exceptional longevity are unknown. We have recently shown that cells from several long-lived bird species, including the budgerigar, are also exceptionally resistant to oxidative damage in vitro. In contrast, cells from a short-lived bird, the Japanese quail, were no more resistant to oxidative damage in vitro than mouse cells. Using cycloheximide to inhibit mRNA translation and actinomycin D to inhibit new transcription, we have shown that both transcription and translation are required for the greater resistance of budgerigar cells to oxidative damage. Based on these results, we have begun looking for changes in gene expression in quail and budgerigar cells in response to oxidative stress. We prepared mRNA from cells incubated in either air or 95% oxygen. Since no microarrays exist for budgerigar, we used a ~ 3000 gene cDNA array containing genes from chicken lymphoid cells. Despite the fact that chickens and budgies are not closely related, we obtained good signals for at least a subset of genes on the array. Genes that are reproducibly over- or under-expressed in budgerigar cells but not in quail cells will be prime candidates for genes that may regulate the response to oxidative damage.
|
||||||||||
|
©Copyright University of Rochester Medical Center, 1999-2006. Disclaimer. For questions or suggestions concerning the content of these pages, contact the URMC Webmaster. |
||||||||||||
